[Info-vax] BOINC for VMS

[Neil Rieck]

On Mar 13, 5:41 pm, Michael Kraemer <M.Krae... at gsi.de> wrote:
> Marc Schlensog schrieb:
>
>
>
> > It was the PCA56, where MVI showed up first. IIRC, MVI is an
> > integer-only extension. If you want to see what DEC had in mind
> > vector-wise for EV8/EV9, check out this PDF:
>
> >www.ptlsim.org/papers/Research/EV9Tarantula.pdf
>
> On paper, LOL.
>
> > Had this ever seen the light of day, it would have blown the
> > competition out of the water, including the performance-per-watt ratio.
>
> It's entirely pointless to bragg about something which never existed,
> and probably never would have, even if DEC had survived.
> DEC couldn't even fix the economical problems associated with the
> lesser Alpha's, let alone with such fantasy chips.

I know we're engaging in arm-chair quarterbacking here, but Alpha was
the fastest chip in its day. But as others have pointed out, it was
very expensive. Why? Well first off, look at all the money they dumped
into building that fab in Hudson Massachusetts (I think it was a
billion). As others had mentioned previously, if DEC had done a deal
with either IBM or Intel to make Alphas, then DEC could have gotten
the price way down while only using up some spare capacity in those
other fabs. But there is the point of view where business types see
the competition as "the enemy" (hey, we don't want "them" to see "our"
designs, and crap like that). Where did it get them in the end? Their
company was carved up and the pieces sold with the largest going to
Compaq which was later merged with HP. Then when HP merged with EDS,
the DEC technology was diluted to irrelevance.

And yet I still clearly remember the electrifying energy seeping
through the DEC campuses in Maynard and Bedford.

Just my two cents worth.

Neil Rieck
Kitchener / Waterloo / Cambridge,
Ontario, Canada.
http://www3.sympatico.ca/n.rieck/